Abstract

The amplified fragment length polymorphism (AFLP) technique was used to examine genetic variation among old and newly emerged individuals of Hyophorbe lagenicaulis (the Round Island bottle palm) on Round Island to assess surviving levels of diversity in the wild population and to evaluate the suitability of old cultivated stocks on Mauritius as a source of seed for reintroduction. The analysis of AFLP data for 48 individuals of H. lagenicaulis (individuals from Round Island and elsewhere), two H. verschaffeltii, two H. vaughanii, one H. amaricaulis and one H. indica yielded 81 variable and six monomorphic bands. Analysis of molecular variance (AMOVA) and Shannon's indices showed a high level of genetic variation within the wild population on Round Island and a smaller amount of genetic variation among cultivated individuals. A neighbor joining analysis resulted in an unrooted network of genetic distances in which the five Hyophorbe spp. were separated and much variation within H. lagenicaulis was recovered. The Round Island populations of H. lagenicaulis contain representatives of the genetic variation found within the species as a whole. However, a few individuals, both wild and cultivated, represent apparently rare AFLP profiles, and thus, if a more representative distribution of genotypes is wanted for the wild population, cultivated sources could be introduced to Round Island from Mauritian gardens and plantings.

INTRODUCTION

Some globally threatened plant taxa have extensive cultivated populations in botanical gardens and nurseries, outnumbering, sometimes by orders of magnitude, the wild stocks. Hyophorbe lagenicaulis (L.H.Bailey) H.E.Moore (the Round Island bottle palm) is such a species, an ornamental palm in the tropics and subtropics that is abundant in cultivation (Maunder et al., 2001a, 2002). Ex situ conservation has been promoted as a means of holding samples of threatened plant species until recovery management and reintroduction are practical (Guerrant, Havens & Maunder, 2004; Namoff et al., 2010). In some respects, H. lagenicaulis already represents a conservation success; the wild population has survived a dramatic population bottleneck and is now regenerating. We use this palm as a case study to quantify the genetic diversity in old garden plantings and to assess whether undocumented collections of threatened species can support species recovery work, including reintroduction.

Hyophorbe lagenicaulis is endemic to the small offshore island, Round Island, off the north-east coast of Mauritius. Round Island is a protected area managed by the National Parks and Conservation Service of the Mauritian government, with technical support from the Mauritius Wildlife Foundation Service, a nongovernmental organization. The importance of H. lagenicaulis and the associated palm woodland habitat is recognized in the International Union for the Conservation of Nature Species Survival Commission (IUCN/SSC) Palm Action Plan (Strahm in Johnson, 1996), and discussed in papers on palm conservation and systematics, for example by Bailey (1942), Moore (1977, 1978), Dransfield (1996) and Maunder et al. (2002). In addition to H. lagenicaulis, the monophyletic genus Hyophorbe Gaertn. consists of another four species: H. verschaffeltii H.Wendl. from Rodrigues, H. indica Gaertn. from La Réunion and H. amaricaulis Mart. and H. vaughanii L.H.Bailey from Mauritius (Lewis & Barboza, 2000; Lewis & Doyle, 2000).

All five Hyophorbe spp. are endemic to the Mascarene Islands, all are listed on the IUCN Red List as Critically Endangered, except H. indica which is Endangered, and all are seriously threatened by extinction in the wild (Maunder et al., 2002; Hilton-Taylor, 2004). Hyophorbe verschaffeltii survives as < 60 individuals in the wild, H. indica survives as c. 500 individuals in the wild, H. vaughanii has only three wild individuals and H. amaricaulis survives as a single individual in the Botanical Garden at Curepipe, Mauritius (Maunder et al., 2002).

Conservation of h. lagenicaulis

Baker (1877), in his Flora of Mauritius and the Seychelles, described H. lagenicaulis (under the synonym H. amaricaulis) as common on Round Island. Travellers' reports (Lloyd, 1846; Johnston, 1894) and photographs from the 1920s (Maunder et al., 2002) indicate that H. lagenicaulis was once abundant on Round Island. However, the wild populations have decreased dramatically over the last 150 years, primarily as a result of introduced feral animals, e.g. goats and rabbits (Strahm, 1996a, b). The woodland cover on Round Island has been lost and replaced with secondary vegetation characterized by open rock faces and erosion gullies with patches of palm thicket, mostly formed by the endemic palm Latania loddigesii Mart. Palm woodland, often referred to as palm savannah, is thought to have been a common coastal habitat on Mauritius, but it survives only, albeit in a degraded form, on offshore islands such as Round Island and Gunner's Quoin (Strahm, 1996c, d). Following the introduction of goats and rabbits to Round Island in the early 19th century, the palm populations were further reduced. Moore (1978) recorded about 15 individuals and, in 1986, only eight adults and 27 seedlings were recorded (North, Bullock & Dulloo, 1994). Field observations by Maunder in 1996 recorded seven mature plants, with only three located in 1998 (Maunder, 1996, pers. comm). The goats were removed from Round Island in 1979 and the rabbits were eradicated in 1986, using an anticoagulant bait followed by shooting, and this resulted, among other things, in a dramatic recovery of the plant cover (North & Bullock, 1986; North et al., 1994; Dransfield, 1996). Subsequently, through natural regeneration, the population of H. lagenicaulis on Round Island has increased and, in 1999, about 350 seedlings were observed. In addition, seeds have been planted and distributed around the island through management projects (Maunder et al., 2002). These seeds originated from botanical gardens and other cultivated populations established with seeds originally collected on Round Island.

As part of a long-term restoration programme for the only wild population of H. lagenicaulis, a new population has been established on Ile aux Aigrettes, a Mauritian Wildlife Foundation Reserve, off the south-east coast of Mauritius (Dulloo et al., 1997). Ile aux Aigrettes was chosen because it is rat free like Round Island, rats being major predators of palm fruits and seedlings. During the 19th century, seeds of H. lagenicaulis were collected from Round Island and brought back to private gardens and botanical gardens on Mauritius. These old plantings, derived from wild collections made at about the time of the population crash, may hold genetic diversity, lost from the wild population, as in the case of English Cypripedium calceolus L. (Orchidaceae; Fay et al., 2009). In addition, the Sir Seewoosagur Ramgoolam Botanic Garden acted as a node for seed distribution, sending large consignments of seed to overseas botanical gardens and horticultural facilities. Many of these batches were distributed under the name H. amaricaulis (Moore & Guého, 1984).

The strategic value of ex situ conservation is often debated (see, for example, Guerrant et al., 2004; Maunder & Byers, 2005; Namoff et al., 2010). However, it is clear that in-country ex situ facilities, such as the Mauritian botanical garden, represent the best opportunity to provide propagules to support species recovery. An overseas facility would present many logistical and legal barriers to repatriation, at the same time as subjecting the managed populations to potentially harmful levels of artificial selection (Stanley Price, Maunder & Soorae, 2004).

The aim of this study was to use amplified fragment length polymorphisms (AFLPs) as molecular markers, as employed in other Indian Ocean palm species (Shapcott et al., 2007) and other taxa (e.g. Andrade et al., 2009; Weeks & Tye, 2009; Bellusci et al., 2010), to examine genetic variation among old and newly emerged individuals of H. lagenicaulis on Round Island, and to estimate the amount of genetic variation, not present in the wild populations on Round Island, but present in nearby cultivated stocks of H. lagenicaulis. Plastid microsatellites were also investigated, but were proven to be invariant (see below). The results of our investigations on variation among individuals can be used to formulate a conservation strategy for H. lagenicaulis, and one way to increase the genetic variation, if appropriate, is the introduction to Round Island of cultivated stocks from botanical gardens.

MATERIAL AND METHODS

Plant materials

The study included a total of 54 individuals of H. lagenicaulis and related Hyophorbe spp. (Table 1); 30 of the H. lagenicaulis individuals were from Round Island, 13 from Mauritius (Old Quarantine Station, Cassis Church, Museum Grounds in Port Louis and Sir Seewoosagur Ramgoolam Botanic Garden), three from the nearby Ile aux Aigrettes, one accession from Montgomery Botanical Center, Florida and one accession from Old Tree, Mombasa, Kenya. Hyophorbe amaricaulis (one sample), H. verschaffeltii (two samples), H. indica (one sample) and H. vaughanii (two samples) were included as outgroups.

Table 1.

Information on individuals of Hyophorbe lagenicaulis, H. verschaffeltii, H. vaughanii, H. amaricaulis and H. indica

Species Individuals Location 
Cultivated plants   
Hyophorbe lagenicaulis 2a, 2b Old Quarantine Station, Mauritius 
Hyophorbe lagenicaulis 3a, 3b Cassis Church, Port Louis, Mauritius 
Hyophorbe lagenicaulis 3c, 3d Museum Grounds, Port Louis, Mauritius 
Hyophorbe lagenicaulis 10, 12–16, 18 Sir Seewoosagur Ramgoolam Botanic Garden, Mauritius 
Hyophorbe lagenicaulis 5–7 Ile aux Aigrettes, islet off the coast of Mauritius 
Hyophorbe lagenicaulis 48 Old Tree, Mombasa, Kenya 
Hyophorbe lagenicaulis 1a Montgomery Tropical Center, Florida, USA 
Round Island plants   
Hyophorbe lagenicaulis 20, 22–37, 39–45, 47, 50–51 Round Island, Mauritius, regenerating plants 
Hyophorbe lagenicaulis 19, 21, 38 Round Island, Mauritius, old plants 
Related species   
Hyophorbe verschaffeltii  17 Sir Seewoosagur Ramgoolam Botanic Garden, Mauritius 
Hyophorbe verschaffeltii  56 National Tropical Botanical Garden, Hawaii 
Hyophorbe vaughanii  9 Black River Gorges National Park, Mauritius 
Hyophorbe vaughanii 413 Florin Reserve, Mauritius 
Hyophorbe indica  55 National Tropical Botanical Garden, Hawaii 
Hyophorbe amaricaulis 415 Curepipe Botanic Garden, Mauritius 
Species Individuals Location 
Cultivated plants   
Hyophorbe lagenicaulis 2a, 2b Old Quarantine Station, Mauritius 
Hyophorbe lagenicaulis 3a, 3b Cassis Church, Port Louis, Mauritius 
Hyophorbe lagenicaulis 3c, 3d Museum Grounds, Port Louis, Mauritius 
Hyophorbe lagenicaulis 10, 12–16, 18 Sir Seewoosagur Ramgoolam Botanic Garden, Mauritius 
Hyophorbe lagenicaulis 5–7 Ile aux Aigrettes, islet off the coast of Mauritius 
Hyophorbe lagenicaulis 48 Old Tree, Mombasa, Kenya 
Hyophorbe lagenicaulis 1a Montgomery Tropical Center, Florida, USA 
Round Island plants   
Hyophorbe lagenicaulis 20, 22–37, 39–45, 47, 50–51 Round Island, Mauritius, regenerating plants 
Hyophorbe lagenicaulis 19, 21, 38 Round Island, Mauritius, old plants 
Related species   
Hyophorbe verschaffeltii  17 Sir Seewoosagur Ramgoolam Botanic Garden, Mauritius 
Hyophorbe verschaffeltii  56 National Tropical Botanical Garden, Hawaii 
Hyophorbe vaughanii  9 Black River Gorges National Park, Mauritius 
Hyophorbe vaughanii 413 Florin Reserve, Mauritius 
Hyophorbe indica  55 National Tropical Botanical Garden, Hawaii 
Hyophorbe amaricaulis 415 Curepipe Botanic Garden, Mauritius 

DNA extraction and AFLP analysis

DNAs were extracted from silica gel-dried leaf material following the 2 × cetyltrimethylammonium bromide (CTAB) protocol of Doyle & Doyle (1987). All samples were purified and concentrated by running 200 µL DNA mixed with 1000 µL PB (binding) buffer (Qiagen) through a Qiagen column, and subsequently eluting with 30 µL TBE (Tris-borate-EDTA) buffer. AFLPs were performed according to the AFLP Plant Mapping Protocol of PE Applied Biosystems Inc. (1996) and the description in Vos et al. (1995). Twelve selective primer combinations were tested on four samples (three H. lagenicaulis and one H. verschaffeltii) for their usefulness. The combination of EcoRI-AAG and MseI-CTG gave most variable bands and the remaining individuals of Hyophorbe were analysed for their AFLP profile with this selective primer combination.

Fragment and data analysis

AFLP products were run on an ABI PRISM 377 automated sequencer (PE Applied Biosystems Inc.) and analysed with Genescan 2.1 and Genotyper 2.0 (both PE Applied Biosystems Inc.). Only amplified fragments with sizes ranging from 50 to 300 base pairs (bp) were scored, because shorter bands could not be accurately sized and longer bands were weak for many of the individuals. A matrix of individuals and band sizes was constructed in Excel (Microsoft). All bands were checked against the computer assigned size and scored as present (1) or absent (0).

The edited binary matrix was copied into PAUP version 4.0b4b for Macintosh (Swofford, 2001) and the data were analysed using the neighbor joining algorithm. The neighbor joining method was chosen to obtain an estimate of genetic distances among individuals because the method does not assume that all lineages have diverged by equal amounts (i.e. the data do not have to be ultrametric). We also carried out a principal coordinates analysis with Jaccard's coefficient using the R Package for Multivariate Analysis version 4.0 (Casgrain & Legendre, 2001; http://www.bio.umontreal.ca/casgrain/en/labo/R/v4/index.html). Jaccard's coefficient was chosen because it excludes shared absences.

The genetic diversity on Round Island relative to the total genetic diversity of cultivated and wild H. lagenicaulis was estimated by the Shannon diversity index calculated in Excel 2000 (Microsoft) and AMOVA using GenAlEx (Peakall & Smouse, 2001). The genetic diversity among the three old trees (accession numbers 19, 21 and 38) on Round Island versus the genetic diversity among the emerging individuals (accession numbers 19–45, 47 and 49–51) was also measured with the Shannon diversity index (calculated in Excel, Microsoft) and AMOVA using GenAlEx (Peakall & Smouse, 2001).

Plastid DNA microsatellites

One set of plastid DNA microsatellite primers (rpl16-190F, 5′-TCATATAGTGACTGTTTCTT-3′; rpl16-18R, 5′-GCTATGCTTAGTGTG TGACTC-3′) was designed on the basis of a matrix of rpl16 plastid DNA sequences for palm species (Asmussen, 1999). The primers amplified a fragment of about 170 bp with a polyA string that varied in length among species, the idea being that this plastid microsatellite area might also vary within species (e.g. Fay & Cowan, 2001; García-Verdugo et al., 2010; Micheneau et al., 2010). PCR was conducted as in Asmussen (1999) and the amplifications were run on an ABI 377 automated sequencer and analysed with Genescan 2.1 and Genotyper 2.0 (PE Applied Biosystems Inc.).

RESULTS

The primer combination of selective bases used in this study, EcoRI-AAG + MseI-CTG (G7), resulted in 86 interpretable bands. Five of these bands were present in all individuals and 10 bands were unique to one accession or were present in all but one individual. The remaining 71 bands were variable in other ways. Within H. lagenicaulis, 80 bands were interpretable, 10 of which were present in all H. lagenicaulis individuals and 70 were variable among H. lagenicaulis individuals. There were no unique bands for Round Island plants, i.e. bands present in all Round Island plants and absent elsewhere. Similarly, no bands were absent in Round Island individuals and present in all cultivated individuals. Twenty-five bands were variable among Round Island plants only, and monomorphic in all individuals outside Round Island.

The result of the neighbor joining analysis showed that, within H. lagenicaulis individuals, one group consisted of entirely Round Island individuals (Fig. 1, individuals 22, 23, 25, 27, 28, 30, 32, 35, 37, 39, 41, 43, 45 and 47). Another group of H. lagenicaulis individuals consisted of all Round Island individuals except for one Mauritian accession [Fig. 1, individuals 2b (the Mauritian accession), 24, 26, 29, 31, 33, 34, 38, 40, 42 and 44]. The middle part of the network was a mixture of H. lagenicaulis individuals from Round Island (19, 20, 21, 36, 50 and 51), Mauritius (2a, 3a, 3c, 3b, 3d, 10, 12, 13, 14, 15, 16 and 18), Ile aux Aigrettes (5, 6 and 7) and two non-Mascarene individuals, Mombasa, Kenya (48) and Montgomery Botanical Center, Florida (1a). Some of the individuals in the middle of the tree were relatively divergent (on long branches); three of these individuals were from Round Island (19, 21 and 36), one from Mauritius (3a) and one from Ile aux Aigrettes (7). Plants 19, 21 and 38 were old reproductive trees, whereas all other collections from Round Island belonged to the generation germinating after goat and rabbit eradication.

Figure 1.

Unrooted neighbor joining tree based on 81 variable amplified fragment length polymorphism (AFLP) bands from 48 individuals of Hyophorbe lagenicaulis, two individuals of H. verschaffeltii, two individuals of H. vaughanii, one accession of H. amaricaulis and one accession of H. indica. The tree is shown as an unrooted network. Accession numbers from Table 1.

Figure 1.

Unrooted neighbor joining tree based on 81 variable amplified fragment length polymorphism (AFLP) bands from 48 individuals of Hyophorbe lagenicaulis, two individuals of H. verschaffeltii, two individuals of H. vaughanii, one accession of H. amaricaulis and one accession of H. indica. The tree is shown as an unrooted network. Accession numbers from Table 1.

The four outgroup species included in the neighbor joining analysis (H. indica, H. amaricaulis, H. vaughanii and H. verschaffeltii) could all be distinguished from H. lagenicaulis, with all the individuals of H. lagenicaulis forming a group distinct from the other Hyophorbe spp. (Fig. 1). Hyophorbe amaricaulis and H. vaughanii were genetically the most similar of the five Hyophorbe spp., and they differed more from H. lagenicaulis than did H. indica and H. verschaffeltii.

In the principal coordinates analysis, the first four coordinates accounted for 21.8%, 13.5%, 8.8% and 5.2% of the variation, respectively (Fig. 2). Again, the four outgroup species were distinguished from H. lagenicaulis.

Figure 2.

The result of a principal coordinates analysis of 81 variable amplified fragment length polymorphism (AFLP) bands from 48 individuals of Hyophorbe lagenicaulis, two individuals of H. verschaffeltii, two individuals of H. vaughanii, one accession of H. amaricaulis and one accession of H. indica. The first two coordinates accounted for 21.8% and 13.5% (=34.3%), respectively, of the variation.

Figure 2.

The result of a principal coordinates analysis of 81 variable amplified fragment length polymorphism (AFLP) bands from 48 individuals of Hyophorbe lagenicaulis, two individuals of H. verschaffeltii, two individuals of H. vaughanii, one accession of H. amaricaulis and one accession of H. indica. The first two coordinates accounted for 21.8% and 13.5% (=34.3%), respectively, of the variation.

The genetic diversity in H. lagenicaulis measured with the Shannon diversity index was 4.04 (Table 2). The total diversity on Round Island was 4.1, and the diversity among old trees on Round Island was 3.85, whereas the diversity among younger, emerging individuals was 4.1.

Table 2.

Results of Shannon's diversity indices for amplified fragment length polymorphism (AFLP) data from Hyophorbe lagenicaulis

AFLPs Shannon's diversity index 
Plants from cultivation (18) 3.7 
All Round Island plants (30) 4.1 
Regenerating plants on Round Island (27) 4.1 
Old plants on Round Island (3) 3.85 
Total for all H. lagenicaulis individuals 4.04 
AFLPs Shannon's diversity index 
Plants from cultivation (18) 3.7 
All Round Island plants (30) 4.1 
Regenerating plants on Round Island (27) 4.1 
Old plants on Round Island (3) 3.85 
Total for all H. lagenicaulis individuals 4.04 

Number of individuals in parentheses.

AMOVA revealed a high within-population diversity (70%) and a smaller diversity among populations (30%). The within-population variation consisted of 30% variation from non-Round Island individuals and 70% variation within the Round Island population (Table 3). The three old plants on Round Island accounted for 6% of the 70% variation among Round Island individuals.

Table 3.

Results of analysis of molecular variance (AMOVA) for amplified fragment length polymorphism (AFLP) data from Hyophorbe lagenicaulis

AFLPs  
Number of bands 80 
Variation among populations 29% 
Variation within populations 71% 
 Variation outside Round Island (18) 30% 
 Variation on Round Island including old trees (30) 70% 
 Variation among old trees (3) on Round Island  6% 
AFLPs  
Number of bands 80 
Variation among populations 29% 
Variation within populations 71% 
 Variation outside Round Island (18) 30% 
 Variation on Round Island including old trees (30) 70% 
 Variation among old trees (3) on Round Island  6% 

Number of individuals in parentheses.

Analysis of the plastid DNA microsatellites resulted in variation among the five Hyophorbe spp., but there was no variation within H. lagenicaulis. The fragment amplified by the plastid DNA microsatellites was 169 bp long in the two individuals of H. verschaffeltii, 170 bp long in all individuals of H. lagenicaulis and 172 bp long in H. indica, H. amaricaulis and the two individuals of H. vaughanii.

DISCUSSION

Genetic variation among the five species ofHyophorbe

All individuals of H. lagenicaulis form a distinct genetic group as expected of a distinct species, and this allows us to draw conclusions on variation within H. lagenicaulis (Figs 1, 2). The AFLP analysis separated the five Hyophorbe spp. from each other, indicating that the genetic distance analysis of AFLP data (neighbor joining analysis) can effectively distinguish them. The relative positions of the taxa are in agreement with the plastid microsatellite study and with an unrooted phylogenetic tree for the genus Hyophorbe based on DNA sequences from the phosphoribulokinase gene (Lewis & Barboza, 2000). The plastid microsatellite investigated in this study did not add any additional variation within H. lagenicaulis.

Genetic variation withinH. lagenicaulis: Round Island individuals versus cultivated plants

All analyses of the AFLP data showed that the diversity within H. lagenicaulis is high: Shannon's diversity index was high for the species (4.04; Table 2), the AMOVA gave a high value relative to other similar studies (Table 3; Maguire, Peakall & Saenger, 2002) and the neighbor joining tree showed a high diversity for the species (Fig. 1).

All analyses of the AFLP data also showed that the diversity of H. lagenicaulis on Round Island is high compared with the diversity of the species. Shannon's diversity index showed that the genetic diversity among individuals on Round Island of 4.1 was as high as the diversity within the species, and higher than that for the plants in cultivation (3.7; Table 2). In general, a Shannon diversity index of 3.5 or more is considered to be high (Kent & Coker, 1992). Results from AMOVA also gave a high within-population diversity on Round Island (70%) relative to the diversity among cultivated plants (30%). The high diversity among H. lagenicaulis on Round Island was also supported by the 25 bands that varied among Round Island plants only. Finally, the neighbor joining analysis showed that the recovering populations of H. lagenicaulis on Round Island accounted for most of the variation found within the species (Fig. 1). Only a few individuals outside Round Island represented additional variation [e.g. 1a (Old Quarantine Station, Mauritius), 3a (Cassis Church, Port Louis), 5 (Museum, Port Louis), 7 (Sir Seewoosagur Ramgoolam Botanic Garden), 12 (Sir Seewoosagur Ramgoolam Botanic Garden), 13 (Sir Seewoosagur Ramgoolam Botanic Garden) and 15 (Ile aux Aigrettes; Fig. 1)].

The genetic diversity found among H. lagenicaulis individuals on Round Island seems to be high in comparison with the genetic diversity found in nonthreatened palm species (e.g. Perera et al., 1998; Cardoso et al., 2000). However, a stringent comparison is not possible because genetic diversity indices, such as Shannon's diversity index, and calculations of within-population variation based on AMOVA were not calculated.

The high level of genetic variation among the regenerating plants relative to the total variation within the species could be explained by a functional seed bank. The recovering plants may represent regeneration from a stored soil seed bank and/or from seed dispersal from the adult trees that were left when the grazing pressure was removed. Studies on the persistence of H. lagenicaulis in a soil seed bank have shown that seeds survive for at least 18 months at 15 °C with no apparent loss in viability (Wood & Pritchard, 2003). This long survival time indicates that fruits from H. lagenicaulis have the physiological potential to survive in dry soil if no predators are present. However, the genetic variation among the three adult plants is high and germinating seeds from these trees may account for most of the variation observed among the regenerating plants. The seed bank and the introduction of seeds may explain the difference, although relatively small, in variation between sampled old and regenerating plants.

Variation on Round Island : old individuals versus recovering plants

The genetic diversity among the three old individuals of H. lagenicaulis (numbers 19, 21 and 38) on Round Island was slightly lower (Shannon diversity index of 3.85) than the diversity among the emerging individuals (Shannon diversity index of 4.1). The relatively small difference in genetic diversity among old and new individuals indicates that the reintroductions of H. lagenicaulis or seeds from recently dead wild trees have had an impact on surviving diversity.

It is not known how many of the eight adult plants contributed to the next generation of seedlings, as we were able to sample only three, namely 19, 21 and 38. However, the sampling in general was representative for the geographical distribution of H. lagenicaulis on Round Island, and also for the morphological and size variation among individuals on Round Island. Thus, the geographical variation and the morphological variation are covered in our study and additional sampling would not be expected to add much additional variation to the already high values of genetic variation found among H. lagenicaulis individuals on Round Island.

The seedlings are clustered in areas with a good layer of soil and some shelter from prevailing winds. The sampled trees are from throughout the island. Palm number 36 has the greatest genetic distance from the majority of the regenerated plants, but this plant does not differ morphologically from the other plants and it is not from a separate area.

Conservation implications

Islands typically contain a higher proportion of threatened species than the mainland (Rieseberg & Swensen, 1996). A number of studies have focused on threatened island plant species, and Rieseberg & Swensen (1996) summarized the problems to be aware of when managing these threatened species, for example, hybridization with related species, clonal growth and implementation of transplantation programmes without first considering genetic factors. These problems are important because the genetic variation within the declining populations can be small. However, the genetic variation found among individuals of H. lagenicaulis on Round Island is high, and the population is not thought to be prone to problems of inbreeding following a population bottleneck. The additional variation outside the Round Island population visualized on the neighbor joining tree could be introduced to Round Island by collecting seeds from these founders for reintroduction (Fig. 1).

A reintroduction of cultivated sources of H. lagenicaulis to Round Island is an important conservation initiative, and such efforts would have considerably more impact than a more generalized forest restoration approach. Round Island supports the last remnant of palm savannah once characteristic of the northern plain of Mauritius, and it has the largest area of native vegetation in Mauritius (North et al., 1994). In addition, Round Island is one of the few elevated tropical islands without introduced rodents and is free of major woody weeds (North et al., 1994). Hyophorbe lagenicaulis is an easily grown palm that flowers and fruits profusely in cultivation (Dransfield, 1996). However, the male flowers open and fall before the female opens, so that there is usually no chance of pollination involving male and female flowers within the same inflorescence. There is the chance of pollen from another inflorescence on the same tree reaching the female flowers, and this happens in cultivated individuals (Dransfield, 1996). It can therefore be anticipated that H. lagenicaulis will regenerate well when reintroduced into its native habitat.

CONCLUSIONS

This study shows that ex situ resources, not necessarily designed as a conservation resource, can provide material to support the recovery of a threatened species. In addition, it shows the critical role of in situ management in releasing a threatened species from a population bottleneck. The Round Island population of H. lagenicaulis is genetically diverse; the measures of genetic diversity (Shannon's diversity index and AMOVA) gave high values for diversity on Round Island and for the species in general, and only a few individuals have AFLP profiles that are relatively different from the general patterns. We recommend that seeds from these sources are utilized for additional planting on Round Island.

This study demonstrates the value of botanical garden collections and historical plantings, and shows that retrospective provenancing can locate plant material of importance for recovery management. Botanical gardens have a proven success in maintaining samples of species close to extinction or extinct in the wild (Maunder et al., 2000, 2001b). Many of the palm plantings sampled on Mauritius were beginning to senesce, and we strongly recommend that these are sampled and seeds from each founder are propagated for planting in secure sites, such as the Sir Seewoosagur Ramgoolam Botanic Garden.

ACKNOWLEDGEMENTS

This work was supported by the Global Environment Facility of the World Bank, the Friends of Kew Threatened Plants Appeal and The Carlsberg Foundation, Denmark. We are grateful for the help and support received from Dr Ehsan Dulloo, Mr Ashok Khadun and Dr Carl Jones from the Mauritian Wildlife Foundation, Mr Yusoof Mungroo, Director of the National Parks and Conservation Service, Mauritius, Mr Raj Rittoo of the Sir Seewoosagur Ramgoolam Botanic Garden and Mr John Hartley of the Durrell Wildlife Conservation Trust. We thank Mr Christian Lange for preparing Figures 1 and 2, and Dr Carl Lewis for providing DNA samples of H. vaughanii and H. amaricaulis.

REFERENCES

Andrade
IM
,
Mayo
SJ
,
van den Berg
C
,
Fay
MF
,
Chester
M
,
Lexer
C
,
Kirkup
D
2009
.
Genetic variation in natural populations of Anthurium sinuatum and A. pentaphyllum var. pentaphyllum (Araceae) from north-east Brazil using AFLP molecular markers
.
Botanical Journal of the Linnean Society
 
159
:
88
105
.
Asmussen
CB
1999
.
Toward a chloroplast DNA phylogeny of the tribe Geonomeae (Palmae)
. In:
Henderson
A
,
Borchsenius
F
eds.
Evolution, variation and classification of palms
 .
The Bronx
:
The New York Botanical Garden Press
,
121
129
.
Bailey
LH
1942
.
Palms of the Mascarenes
.
Gentes Herbarum
 
6
:
50
104
.
Baker
JG
1877
.
Flora of Mauritius and the Seychelles
 .
London
:
L. Reeve & Co
.
Bellusci
F
,
Musacchio
A
,
Palermo
AM
,
Pellegrino
G
2010
.
Exploring the history of Serapias politisii (Orchidaceae), a narrow endemic on the opposite coasts of the Otranto channel: insights from molecular investigations
.
Botanical Journal of the Linnean Society
 
162
:
572
580
.
Cardoso
SRS
,
Eloy
NB
,
Provan
J
,
Cardoso
MA
,
Ferreira
PCG
2000
.
Genetic differentiation of Euterpe edulis Mart. populations estimated by AFLP analysis
.
Molecular Ecology
 
9
:
1753
1760
.
Casgrain
P
,
Legendre
P
2001
.
R 4.0. The R Package for multidimensional and spatial analysis
 . http://www.bio.umontreal.ca/casgrain/en/labo/R/v4/index.html. Accessed August 2011.
Doyle
JD
,
Doyle
JL
1987
.
A rapid DNA isolation procedure from small quantities of fresh leaf tissue
.
Phytochemical Bulletin, Botanical Society of America
 
19
:
11
15
.
Dransfield
J
1996
.
Hyophorbe lagenicaulis
.
Curtis's Botanical Magazine
 
13
:
186
190
.
Dulloo
ME
,
Verburg
SS
,
Paul
S
,
Green
SE
,
de Boucherville Baisac
P
,
Jones
C
1997
.
Ile aux Aigrettes management plan 1997–2000
 . Mauritian Wildlife Foundation Technical Series 1/97.
Port Louis
:
Mauritian Wildlife Foundation
.
Fay
MF
,
Bone
R
,
Cook
P
,
Kahandawala
I
,
Greensmith
J
,
Harris
S
,
Pedersen
,
Ingrouille
MJ
,
Lexer
C
2009
.
Genetic diversity in Cypripedium calceolus (Orchidaceae) with a focus on northwestern Europe, as revealed by plastid DNA length polymorphisms
.
Annals of Botany
 
104
:
517
525
.
Fay
MF
,
Cowan
RS
2001
.
Plastid microsatellites in Cypripedium calceolus (Orchidaceae): genetic fingerprints from herbarium specimens
.
Lindleyana
 
16
:
151
156
.
García-Verdugo
C
,
Forrest
AD
,
Balaguer
L
,
Fay
MF
,
Vargas
P
2010
.
Parallel evolution of insular Olea europaea subspecies based on geographical structuring of plastid DNA variation and phenotypic similarity in leaf traits
.
Botanical Journal of the Linnean Society
 
162
:
54
63
.
Guerrant
EO
,
Havens
K
,
Maunder
M
eds.
2004
.
Ex situ plant conservation: supporting species survival in the wild
 .
Washington, DC
:
Island Press
.
Hilton-Taylor
C
(compiler).
2004
.
2000 IUCN red list of threatened species
 .
Gland and Cambridge
:
IUCN
(http://www.redlist.org). Accessed August 2011.
Johnson
D
ed.
1996
.
Palms, their conservation and sustained utilization. IUCN/SSC Palm Specialist Group
 .
Gland and Cambridge
:
IUCN
.
Johnston
HH
1894
.
Report on the flora of Round Island, Mauritius
.
Transactions and Proceedings of the Botanical Society of Edinburgh
 
20
:
237
264
.
Kent
M
,
Coker
P
1992
.
Vegetation description and analysis: a practical approach
 .
London
:
Belhaven Press
.
Lewis
CE
,
Barboza
NM
2000
.
Identity of the Hyophorbe palms at the Botanical Garden of Cienfuegos, Cuba
.
Palms
 
44
:
93
97
.
Lewis
CE
,
Doyle
JJ
2000
.
Low-copy nuclear genes for phylogenetic research in the palm family
.
American Journal of Botany, Abstract
 
87
:
139
.
Lloyd
J
1846
.
Account of a voyage to Round Island and Serpent Island in December 1844
.
Transactions and Proceedings of the Natural History Society of Mauritius
 
1845
:
154
162
.
Maguire
TL
,
Peakall
R
,
Saenger
P
2002
.
Comparative analysis of genetic diversity in the mangrove species Avicennia marina (Forsk.) Vierh. (Avicenniaceae) detected by AFLPs and SSRs
.
Theoretical and Applied Genetics
 
104
:
388
398
.
Maunder
M
,
Byers
O
2005
.
The IUCN technical guidelines on the management of ex situ populations for conservation: reflecting major changes in the application of ex situ conservation
.
Oryx
 
39
:
95
98
.
Maunder
M
,
Cowan
RS
,
Stranc
P
,
Fay
MF
2001b
.
The genetic status and conservation management of two cultivated bulb species extinct in the wild: Tecophilaea cyanocrocus (Chile) and Tulipa sprengeri (Turkey)
.
Conservation Genetics
 
2
:
193
201
.
Maunder
M
,
Culham
A
,
Alden
B
,
Zizka
G
,
Orliac
C
,
Lobin
W
,
Bordeu
A
,
Ramirez
JM
,
Glissmann-Gough
S
2000
.
Conservation of the toromiro tree: case study in the management of a plant extinct in the wild
.
Conservation Biology
 
14
:
1341
1350
.
Maunder
M
,
Lyte
B
,
Dransfield
J
,
Baker
W
2001a
.
The conservation value of botanic garden palm collections
.
Biological Conservation
 
98
:
259
271
.
Maunder
M
,
Page
W
,
Mauremootoo
J
,
Payendee
R
,
Mungroo
Y
,
Maljkovic
A
,
Vericel
C
,
Lyte
B
2002
.
The decline and conservation management of the highly threatened endemic palms of the Mascarene Islands
.
Oryx
 
36
:
56
65
.
Micheneau
C
,
Duffy
KJ
,
Smith
RJ
,
Stevens
LJ
,
Stout
JC
,
Civeyrel
L
,
Cowan
RS
,
Fay
MF
2010
.
Plastid microsatellites for the study of genetic variability in the widespread Cephalanthera longifolia, C. damasonium and C. rubra (Neottieae, Orchidaceae), and cross amplification in other Cephalanthera species
.
Botanical Journal of the Linnean Society
 
163
:
181
193
.
Moore
HE
1977
.
Endangerment at the specific and generic levels in palms
. In:
Prance
GT
,
Elias
TS
eds.
Extinction is forever
 .
The Bronx
:
New York Botanical Garden Press
,
267
282
.
Moore
HE
1978
.
The genus Hyophorbe (Palmae)
.
Gentes Herbarum
 
11
:
212
245
.
Moore
HE
,
Guého
LJ
1984
.
Palmiers
. In:
Bosser
J
,
Cadet
T
,
Guého
J
,
Marais
W
eds.
Flore des Mascareignes, fascicule. 189
 .
Reduit
:
Sugar Industry Research Institute
;
Paris
:
ORSTOM
;
Kew
:
Royal Botanic Gardens
,
1
34
.
Namoff
S
,
Husby
CE
,
Francisco-Ortega
J
,
Noblick
LR
,
Lewis
CE
,
Griffith
MP
2010
.
How well does a botanical garden collection of a rare palm capture the genetic variation in a wild population?
Biological Conservation
 
143
:
1110
1117
.
North
SG
,
Bullock
DJ
1986
.
Changes in the vegetation and populations of introduced mammals on Round Island and Gunner's Quoin, Mauritius
.
Biological Conservation
 
37
:
99
118
.
North
SG
,
Bullock
DJ
,
Dulloo
ME
1994
.
Changes in the vegetation and reptile populations of Round Island, Mauritius, following eradication of rabbits
.
Biological Conservation
 
67
:
21
28
.
PE Applied Biosystems Inc.
1996
.
AFLP plant mapping protocol
 .
Carlsbad, CA
:
PE Applied Biosystems Inc
.
Peakall
R
,
Smouse
PE
2001
.
Genalex V5: genetic analysis in Excel. Population genetic software for teaching and research
 .
Canberra
:
Australian National University
, http://www.anu.edu.au/BoZo/GenAlEx/. Accessed August 2011.
Perera
L
,
Russell
JR
,
Provan
J
,
McNicol
JW
,
Powell
W
1998
.
Evaluating genetic relationships between indigenous coconut (Cocos nucifera L.) accessions from Sri Lanka by means of AFLP profiling
.
Theoretical and Applied Genetics
 
96
:
545
550
.
Rieseberg
LH
,
Swensen
SM
1996
.
Conservation genetics of endangered island plants
. In:
Avise
JC
,
Hamrick
JL
eds.
Conservation genetics. Case histories from nature
 .
London
:
Chapman & Hall
,
305
334
.
Shapcott
A
,
Rakotoarinivo
M
,
Smith
RJ
,
Lysaková
G
,
Fay
MF
,
Dransfield
J
2007
.
Can we bring Madagascar's critically endangered palms back from the brink? Using an understanding of genetics and ecology to guide a conservation and recovery programme for the iconic and critically endangered palm Beccariophoenix madagascariensis
.
Botanical Journal of the Linnean Society
 
154
:
589
608
.
Stanley Price
MR
,
Maunder
M
,
Soorae
PS
2004
.
Ex situ support to the conservation of wild populations and habitats: lessons from zoos and opportunities for botanic gardens
. In:
Guerrant
EO
,
Havens
K
,
Maunder
M
eds.
Ex situ plant conservation: supporting species survival in the wild
 .
Washington, DC
:
Island Press
,
84
111
.
Strahm
W
1996a
.
Botanical history of the Mascarene Islands
.
Curtis's Botanical Magazine
 
13
:
217
228
.
Strahm
W
1996b
.
Conservation of the flora of the Mascarene Islands
.
Curtis's Botanical Magazine
 
13
:
228
237
.
Strahm
W
1996c
.
Mascarene Islands – an introduction
.
Curtis's Botanical Magazine
 
13
:
182
185
.
Strahm
W
1996d
.
The vegetation of the Mascarene Islands
.
Curtis's Botanical Magazine
 
13
:
214
217
.
Swofford
DL
2001
.
Paup*. Phylogenetic analysis using parsimony (*and other methods), version 4
 .
Sunderland, MA
:
Sinauer Associates
.
Vos
P
,
Hogers
R
,
Bleeker
M
,
Reijans
M
,
van de Lee
T
,
Hornes
M
,
Frijters
A
,
Pot
J
,
Peleman
J
,
Kuiper
M
,
Zabeau
M
1995
.
AFLP: a new technique for DNA fingerprinting
.
Nucleic Acids Research
 
23
:
4407
4414
.
Weeks
A
,
Tye
A
2009
.
Phylogeography of palo santo trees (Bursera graveolens and Bursera malacophylla; Burseraceae) in the Galápagos archipelago
.
Botanical Journal of the Linnean Society
 
161
:
396
410
.
Wood
CB
,
Pritchard
HW
2003
.
Germination characteristics of fresh and dried Hyophorbe lagenicaulis seeds
.
Palms
 
47
:
45
50
.

Author notes

Current address: Al Ain Wildlife Parks and Resort, PO Box 1204, Al Ain, United Arab Emirates